PARCHED POWER: WATER DEMANDS, RISKS, AND OPPORTUNITIES FOR INDIA S POWER SECTOR

Transcription

1 WORKING PAPER PARCHED POWER: WATER DEMANDS, RISKS, AND OPPORTUNITIES FOR INDIA S POWER SECTOR TIANYI LUO, DEEPAK KRISHNAN, AND SHREYAN SEN EXECUTIVE SUMMARY Highlights India s thermal power sector is very dependent on water and has been suffering from water shortages, losing a substantial part of its generation growth every year since 213. Most of the country s existing plants are likely to experience an increased level of water competition by 23. Fourteen of India s top 2 largest thermal power utility companies have experienced water shortage related disruptions at least once between 213 and 216, losing more than $1.4 billion in total potential revenue. Water consumption from India s thermal power generation rose steadily every year between 211 and 216 but would stay below its 216 level by 227 if the country s most ambitious renewable goals are successfully achieved and the notified stringent water regulations implemented. This study provides a first-cut assessment of the water risks associated with India s thermal power sector, leveraging a new plant-level database with information on cooling technology, source water type, water withdrawal and consumption, and actual generation for all thermal utilities in India. The Ministry of Power, Government of India, should mandate that power plants monitor and disclose water withdrawal and discharge data, create guidelines and policy incentives to drive better performance in managing water use and risks, and prioritize solar photovoltaic (PV) and wind projects when possible. CONTENTS Executive Summary Introduction Data and Methodology Water Use, Risks, and Opportunities for India s Thermal Power Sector Water Demands Risks Opportunities Recommendations Limitations...34 Appendix...35 References...38 Working Papers contain preliminary research, analysis, findings, and recommendations. They are circulated to stimulate timely discussion and critical feedback, and to influence ongoing debate on emerging issues. Working papers may eventually be published in another form and their content may be revised. Suggested Citation: Luo, Tianyi, Deepak Krishnan, and Shreyan Sen Parched Power: Water Demands, Risks, and Opportunities for India s Power Sector. Working Paper. Washington, DC: World Resources Institute. Available online at WORKING PAPER January 218 1

2 Context India s demand for water will continue to grow, despite being an already water-stressed nation. Freshwater resources are already scarce in most parts of India (Shiao et al. 215). As India s economy is projected to double by 23 (PwC 217), the country s water demand is also expected to grow significantly across sectors (CWC 215). The power sector in India is very dependent on water and has been suffering from droughts and water shortages. More than 8 percent of India s electricity is generated from thermal (fossil fuel, biomass, nuclear, and concentrated solar) power plants (CEA 217) that rely significantly on water for cooling. Another 1 percent of electricity is generated from hydroelectric plants, which depend on water completely. Thermal power plants have been forced to shut down due to inaccessibility of cooling water, losing tens of terawatt-hours of electricity generation in recent years (Luo 217). This paper aims to help decision-makers understand the magnitude of water issues for the thermal power sector in India with quantitative evidence. There is a significant data gap in power plant water use in India. The authors used data science techniques and innovative methodologies and developed a comprehensive plant-level geodatabase on water withdrawal and consumption for India s thermal power sector, making a first-cut attempt to fill the data gap. Combined with information on power generation, water risks, and future projections of energy and water demand, this paper quantifies the Indian thermal power sector s water demand, assesses its exposure to water stress, and evaluates opportunities for reducing water requirements while supporting power growth for the future. Box ES-1 Definitions of Water Withdrawal and Consumption WATER WITHDRAWAL: the total amount of water that is diverted from a water source (e.g., surface water, groundwater) for use. WATER CONSUMPTION: the portion of water withdrawal that is not returned to the original water source after being withdrawn. Key Findings Almost 9 percent of India s thermal power generation depends on freshwater for cooling. In 216, thermal (fossil and nuclear) electricity accounted for more than 83 percent of India s total utility power generation (CEA 217). More than 8 percent of the total thermal generation was cooled by freshwater recirculating systems, as shown in Figure ES-1. Freshwater oncethrough systems are the second most common cooling technology in India, accounting for about 7 percent of total thermal generation in 216. Freshwater consumption from Indian thermal utilities increased by 43 percent from 211 to 216, while withdrawals stayed fairly stable. The increase in consumption is due to the steady growth in electricity generation, as illustrated in Figure ES-2, and an increased share of electricity generated by plants with recirculating cooling systems. Stable water withdrawals during the period reflect that no new freshwater oncethrough cooled power plants were built after 211. Although water withdrawals have not increased, the substantial increase in freshwater consumption means there is reduced freshwater available to other sectors. Figure ES-1 India s Thermal Utility Power Generation Distribution by Water Source and Cooling Technology In 216, almost 9% of India's thermal utility power generation used freshwater for cooling 8% 6% 5% 2% 7% Seawater Once-Through Seawater Recirculating Freshwater Dry Cooling Freshwater Once-Through Freshwater Recirculating Source: Reig (213). Source: WRI authors 2

3 Parched Power: Water Demands, Risks, and Opportunities for India s Power Sector Figure ES-2 India s Annual Thermal Utility Generation, Freshwater Consumption, and Withdrawal between 211 and 216 Thermal utility generation grew by 4% between 211 and 216 Freshwater consumption increased by 43% between 211 and 216 Freshwater withdrawal stayed relatively stable between 211 and 216 Terawatt-Hours 1,2 1, Billion Cubic Meters Billion Cubic Meters Sources: WRI authors; CEA (217) India lost about 14 terawatt-hours of thermal power generation due to water shortages in 216, canceling out more than 2 percent of growth in the country s total electricity generation from 215. Between 213 and 216, as shown in Figure ES-3, 61 percent of the time programmed daily thermal generation targets couldn t be met due to forced power plant outages, which included equipment failure, fuel shortages, water shortages, and other factors. Based on the Daily Outage Reports disclosed by the Central Electricity Authority (CEA) between 213 and 216, water shortage is the fifth most common reason for forced outages of Indian thermal power plants and caused almost 2 percent of all outages in terms of potential generation. Among all of India s freshwater-cooled thermal utilities, 39 percent of the capacity is installed in high water-stress regions. That capacity generated 34 percent of the total freshwater-cooled thermal power generation in 216. Water stress is the ratio of total water Figure ES-3 India s Daily Programmed and Actual Generation from Thermal Power Utilities between 213 and % of the time, programmed thermal generation couldn't be delivered due to forced power plant outages, including equipment failure, fuel shortages, water shortages, and so on Daily Generation (GWh) 3, 2,5 2, 1,5 Programmed Thermal Generation Actual Thermal Generation Thermal Generation Surplus Thermal Generation Deficit Differences Between Programmed and Actual Generation (GWh) Source: Data from CEA, compiled and analyzed by WRI authors. WORKING PAPER January 218 3

4 withdrawal over available supply (Gassert et al. 214). High water stress indicates a high level of competition in water use. Figure ES-4 is a water-stress map with all freshwater-cooled thermal power utilities in India. Freshwater-cooled thermal power plants that are located in high water-stress areas have a 21 percent lower average capacity factor, compared to the ones in low and medium water-stress areas. Among India s 19 ultra large freshwater-cooled plants (with an installed capacity over two gigawatts), 16 are located in low-and medium water-stress regions. Furthermore, we controlled the comparison analysis by unit age, fuel type, and plant capacity and observed the same trend in almost every control group: Plants in high-stress areas have a lower average capacity factor than those in low and medium water-stress areas. Some of the most disruptive water shortages occurred in India s most water-abundant area. We also found that, even in water-abundant or low water-stress regions, thermal plants can still face water shortage related risks during droughts or when monsoons are delayed. Some of those plants for example, Farakka, Raichur, and Tiroda experienced significant, if not the biggest, disruptions in generation caused by water shortages. Fourteen of India s largest thermal power utility companies have experienced water shortagerelated disruptions at least once between 213 and 216, losing over $1.4 billion in total potential revenue from the sale of power, and are likely to continue facing the problem as water competition intensifies in the future. In 216, nine companies had water-related shutdown records for 12 of their plants, and together lost more than $614 million in potential revenue, accounting for about 2.3 percent of their total revenue from the sale of power in 216. For assessing companies exposure to water risks, we benchmarked India s 2 largest thermal utility companies against four waterrelated metrics, as shown in Table ES-1. Freshwater consumption from India s thermal power generation would stay below its 216 level by 227 if the country s most aggressive renewable targets are achieved and the notified stringent power-sector water regulations implemented. We analyzed two scenarios scenario 1 (developed by CEA) and scenario 2 (developed by WRI authors based on CEA s draft national electricity plan) for the year 227, as well as the notified power sector water regulations by the Ministry of Environment, Forest, and Climate Change (MOEFCC). We found that, for the thermal power sector under scenario 2, despite a more than 6 percent projected increase in electricity generation, freshwater consumption would stay below the 216 level, and water withdrawals would be reduced significantly by more than 12 billion cubic meters. However, even maintaining 216 water consumption levels implies continued risk of electricity outages, competition with other rapidly growing sectors, and increased variability in local water supplies due to climate change. There is a huge data gap in water withdrawal and consumption information for India s power sector. Our research attempts to fill the gap, but the limitations of our method and data cannot substitute for actual ground-level measurement and monitoring. In fact, our estimates of India s total thermal power sector water use could be on the lower end. Additionally, our benchmarking of utility companies does not capture corporate water management practices and technological innovations. Terms such as water withdrawal and water consumption are used interchangeably by India s power sector. For example, MOEFCC s notified regulations use the term specific water consumption, when referring to what is conventionally called water withdrawal. In fact, power plants in India currently only measure water withdrawal, not consumption. The lack of standardization in terminology could create confusion in water-use monitoring and accounting. Recommendations The Ministry of Power, Government of India, should mandate that power plants start monitoring and disclosing water withdrawal and discharge data, leveraging its existing daily reporting system. Currently, there is a significant data gap in power plant water withdrawal and consumption information in India. Unlike the detailed generation and capacity data one can easily find about power plants, water-related data are scarce and difficult to find. Mandating monitoring and disclosure will help promote transparency and accountability in how the power sector manages water resources and build the foundation for assessing risks and measuring progress. 4

5 Parched Power: Water Demands, Risks, and Opportunities for India s Power Sector Figure ES-4 India s Freshwater-Cooled Thermal Utilities Mapped against Baseline Water Stress and Distribution in Installed Capacity by Water Stress Level by State Delhi Freshwater Dry Cooling Freshwater Once-Through Ahmedabad Kolkata Freshwater Recirculating Low (<1%) Low to Medium (1 2%) Mumbai Medium to High (2 4%) Hyderabad High (4 8%) Extremely High (>8%) Arid & Low Water-Use Bangalore Chennai Maharashtra Uttar Pradesh Gujarat Madhya Pradesh Chhattisgarh West Bengal Andhra Pradesh Rajasthan Odisha Tamil Nadu Jharkhand Installed Capacity (GW) Karnataka Haryana Telangana Punjab Bihar Delhi Tripura Assam Kerala Note: Symbol size reflects the power plant s relative installed capacity. Source: WRI authors. Disclaimer: This map is for illustrative purposes and does not imply the expression of any opinion on the part of WRI concerning the legal status of any country or territory or concerning the delimitation of frontiers or boundaries. WORKING PAPER January 218 5

6 Table ES-1 Water Dependency and Risk Exposure Benchmarking for India s Largest Thermal Utility Companies as of December 216 COMPANY TOTAL THERMAL CAPACITY (GW) FRESHWATER WITHDRAWAL INTENSITY (M³/MWH) NO. OF ASSETS THAT HAD AT LEAST ONE WATER SHORTAGE INDUCED SHUTDOWN RECORDED BETWEEN 213 AND 216 % REVENUE GENERATED IN HIGH WATER- STRESS AREA PROJECTED CHANGE IN FUTURE WATER-USE COMPETITION WITH OTHER WATERSHED STAKEHOLDERS NTPC % 9.9% Adani Power % 3.% MSEB Holding Co % 27.7% Damodar Valley Corp % 21.% Reliance % 13.8% Tata Group % 14.2% Gujarat State Elec. Corp % 23.1% Nuclear Power Corp % 1.4% Uttar Pradesh RV % 8.1% Tamil Nadu Gen. & Dist. Corp % 5.7% Rajasthan RVUN % 5.8% West Bengal Power Dev. Corp % 8.2% Andhra Pradesh Power Gen. Corp % 5.9% MP Power % 11.5% Essar Energy % 13.7% GMR Group % 15.3% Karnataka Power Corp % 15.8% Haryana Power Gen Co % -5.2% Vedanta Resources % 9.% Torrent Power % 18.5% Top 25% Upper middle 25% Lower middle 25% Bottom 25% Notes: Only thermal plants are included in the benchmarking exercise. Capacity data are from the Platts World Electric Power Plant database and might have small discrepancies compared to data disclosed by CEA or other sources. Color codes represent relative performance between companies within each metric. Source: WRI authors. 6

7 Parched Power: Water Demands, Risks, and Opportunities for India s Power Sector Reporting on water data monitoring and disclosure for power plants should be standardized. Unlike greenhouse gas emissions, there is no widely recognized guideline or standard on how power plants should account for and report on their water usage. For example, terms such as water withdrawal and water consumption are used interchangeably by India s power sector. The lack of standardization of terminology and calculation methodologies makes it difficult for utilities to monitor and disclose their water data, discouraging them from reporting, and weakening the comparability and usefulness of the data. A standardized thermal power sector water data reporting method would provide consistency and clarity, help policymakers develop and implement specific water conservation regulations, and guide utility companies in monitoring and disclosing their water performance. The Ministry of Power, Government of India, should set power sector water performance benchmarking guidelines and create policy guidelines and incentives for better performers. Water dependency and risk exposure vary greatly among companies. Some are more freshwater efficient and have less environmental impact than others. Both public and private power utility companies water performance should be benchmarked with standardized monitored data and corporate disclosure. Utilities that are better at managing water and controlling risks have lower chances of disruptions in their services during extreme drought and should be recognized and rewarded for their effort and ability to provide greater stability and more reliable services through regulations and incentives created by the Ministry of Power. Thermal power utility companies should investigate and assess their water-related risks to identify assets at risk and invest in risk-mitigation or reduction efforts to ensure business continuity and to prepare for future uncertainty. Some Indian power plants have experienced significant, if not the biggest, disruptions in electricity generation, caused purely by water shortages in recent years. Conducting a portfolio-level assessment on water dependency and risk exposure is the key to understanding risks, prioritizing resources, and informing effective mitigation strategies. Additionally, climate change impacts and economic growth will add additional challenges, making it crucial to reassess watershed hydrology at the individual power plant level, including quantifying potential changes in drought probabilities to inform contingency plans and long-term business development planning. Public and private sector investors should assess their investment portfolios exposure to water risks, identify highly exposed companies, and urgently engage those companies in promoting better water management practices and reducing such risks. Fourteen of India s 2 largest thermal utility companies experienced water shortage induced power plant shutdowns at least once between 213 and 216, losing more than $1.4 billion in total in potential revenue from the sale of power. Additionally, these companies are likely to see an increase in water-use competition by 23 and therefore would continue experiencing water-related disruptions if they continue business as usual. Investors (including public financial institutions like development banks) should leverage water risk assessment to engage with companies in which they invest, further identifying company strategies to address water scarcity issues, and ultimately pushing companies to be more sustainable and socially responsible, thereby benefiting both people and the environment. The Government of India should keep working toward meeting its ambitious renewable goals and should prioritize solar PV and wind projects when possible, to scale up power production while reducing the power sector s exposure to waterrelated risks. Under the scenario 2, by 227, India s power sector (hydro excluded) would see a 76 percent decrease in water withdrawal intensity; more than 32 percent of that reduction is driven by the country s power mix shifting toward more solar PV and wind. Water consumption intensity would decrease by about 25 percent; almost 98 percent of that reduction would be driven by the power mix shift. Compared to cooling technology advancement or plant efficiency enhancement, transitioning to more solar PV and wind generation is the only pathway at scale that can cut back both water withdrawal and consumption while sustaining growth in power generation. This is essential to reducing not only the power sector s water dependency and exposure to water risks, but also its impact on the ecosystem and other water users at the national scale. WORKING PAPER January 218 7

8 INTRODUCTION Since 2, India has been making great progress in expanding its power supply to meet the country s steadily growing demand (IEA 216; BP 217). Access to electricity improved from 6 percent of the population in 2 to 79 percent of the population in 214 (World Bank 217). More than three quarters of India s electricity is generated from thermal power plants (CEA 217), which rely significantly on freshwater needs for cooling purposes. India is one of most water-stressed countries in the world, and freshwater resources are scarce in most parts of the country (Shiao et al. 215). India s thermal power sector has suffered from water shortages and lost tens of terawatt-hours of generation and billions of dollars of revenue in the past few years (Luo 217), posing threats to both Indian society and companies. Additionally, as India s economy and demand for power continues to expand, the country s water demand also is expected to grow significantly across sectors by 25 (CWC 215). The competition for available water is only going to become more intense (MWR 212). To address these issues, the Government of India has developed several promising plans, including capping specific water consumption for thermal power plants (MOEFCC 215b), requiring certain plants to use treated wastewater for cooling (MOP 216), setting ambitious targets for renewable energies that are almost waterindependent (MOEFCC 215a), and proposing new water allocation and management principles (MOWR 212). It is our goal to help decision-makers understand the magnitude of water issues in the power sector in India and to provide information that the country can use for informed decision-making in the future. To do this, WRI authors used data science techniques and innovative methodologies and developed a comprehensive plantlevel geodatabase on water withdrawal and consumption for India s power sector. Combined with water risk data from WRI Aqueduct and power projections from India s Central Electricity Authority (CEA), we quantified the Indian power sector s water demand, assessed its exposure to water risks, and evaluated opportunities for reducing water demand while supporting power growth for the future. DATA AND METHODOLOGY Estimates for water withdrawal and consumption attributable to the Indian energy sector vary considerably (Chaturvedi et al. 217; IEA 216; Bhattacharya and Mitra 213). The primary reason is the huge data gap in power plant cooling type and water usages, specifically, the lack of information on power sector cooling technology shares and Indian specific power plant water withdrawal and consumption factors. In this study, WRI used data science techniques and innovative methodologies to fill the data gap, and produced the most up-to-date water-related data at plant level for India s power sector. Here are the three most valuable and unique aspects about the data we developed and used in this study: By applying a recently developed WRI methodology that draws from high resolution satellite images to identify power plant cooling technology, we created a plant-level geodatabase on cooling technology and water sources that represents the on-the-ground situation in 216 (Luo et al. 218). We collected four years of daily generation reports and eight years of monthly reports from India s Central Electricity Authority (CEA) for each plant for which CEA discloses data and structured this information into machine-readable data sets. We developed a data imputation model to fill in missing values of cooling and source water types and capacity factors, leveraging a random forest-based machine learning algorithm trained with observations from the previous two steps. Please refer to the appendix for details on the imputation method and performance. The newly developed data made it possible to: provide a more accurate estimate of annual freshwater withdrawal and consumption by India s power sector as a whole from 211 through 216; develop a water withdrawal and consumption time series for individual power plants; analyze power sector water-use behavior both spatially and temporally at any scale, including plant, city, watershed, state, and national; and evaluate future power sector (hydro excluded) water demand in combination with power mix projections. 8

9 Parched Power: Water Demands, Risks, and Opportunities for India s Power Sector Scope In this study, unless stated otherwise, we focus on thermal power utilities, both government/privately owned utilities and private power producers, and include all plants fueled by coal, nuclear, oil, gas, biomass, and concentrated solar power. Unless otherwise noted, we excluded from all our analyses captive plants, power plants that are owned and operated by industrial and commercial energy users for their own energy consumption. We differentiate between water withdrawal and consumption in power plant water use. Withdrawal is the amount of water diverted from a water source. Some portion of the withdrawal is evaporated or consumed during the generation process, which is defined and measured as consumption. The quantification of water withdrawal and consumption is bounded within power plants themselves and for power generation only. Water demand associated with upstream and downstream energy sector activities like fuel production and electricity consumption is not included in our analysis. Cooling technologies are grouped into three generic categories: once-through, recirculating, and dry cooling. Please refer to Luo et al. (218) for details on cooling technology definition and system diagrams. In this study, we focus on water shortage related risks only. Other types of risks associated with water for example, water temperature, water quality, and flooding are beyond the scope of this paper. Data Sources To fill the data gap to the highest possible degree, we harnessed the best available data for each portion of our model and compiled them into a master geodatabase for our analysis. Table 1 is a list of source data we drew upon to build the plant-level water withdrawal and consumption database and conduct the assessment on water use, risks, and opportunities for India s power sector. Methodology For developing the plant-level water withdrawal and consumption database, we applied the following six-step approach: Plant inventory development. The Platts database was used as the inventory from which we developed a full list of 478 thermal power plants that were in operation in India as of December 216. The inventory database includes information on plant name, installed capacity, fuel type, company name, parent company name, installation year, city and state, and business type at the generating unit level. Power plant geocoding. Through a public power plant geolocation database for example, Global Energy Observatory and others and Internet research, we geolocated and validated on Google Maps the exact latitude and longitude of 358 plants covering about 24 GW, which accounts for almost 99 percent of the total capacity of all thermal power utilities, according to Platts data. Cooling technology identification. We analyzed satellite images for all geolocated utilities and identified water source and cooling type by applying the methodology from Luo et al. (218). Assignment of water withdrawal and consumption factors. We used Indian specific water withdrawal and consumption factors from Chaturvedi et al. (217), when available, and assigned factors to each plant by its cooling and fuel type following the method described in Luo et al. (218). Detailed water factors may be found in the appendix. Annual capacity factor calculation. Eight years of plant-level annual capacity factors are calculated by averaging plant-level monthly capacity factors within each year, estimated with monthly generation and capacity data from CEA. The estimated plant-level annual capacity factors are then matched back to the plant inventory database (step a). In this way, we matched 93.8 percent of the total capacity in the inventory. Missing data imputation. A random forest-based machine learning algorithm was used in imputing missing values of cooling and source water type for the plants we were not able to geolocate and capacity factors for the plants for which we could not find generation data from CEA. Details about our imputation model may be found in the appendix. WORKING PAPER January 218 9

10 Table 1 A List of Source Data Used in the Study DATA TIME FRAME AVAILABILITY SOURCES Plant and regional-level daily generation data January 213 December 216 Public CEA data compiled by WRI authors Plant-level monthly generation and capacity data January 28 December 216 Public CEA data compiled by WRI authors Unit-level rate of sale of power data Fiscal year 215 Public CEA data compiled by WRI authors Unit-level daily outage data January 213 December 216 Public CEA data compiled by WRI authors Country-specific water withdrawal and consumption factors by fuel and cooling type Not applicable Public Chaturvedi et al. (217); CWR/IRENA (216); Bhattacharya and Mitra (213); NREL (211) Unit-level capacity, built year, operating status, fuel, business type, and ownership data Calendar year 216 Proprietary The Platts World Electric Power Plants Database Catchment-level current water supply, demand, and stress data Calendar year 21 Public Gassert et al. (214) Catchment-level projected future water supply, demand, and stress data Calendar year 23 Public Luck et al. (215) National-level entire power sector generation and capacity data Calendar year 214 Public Courtesy of IRENA (International Renewable Energy Agency) National-level projected future power mix, generation, and capacity Calendar year 227 Public Courtesy of CEA Limitations In this paper, neither water quality or water temperature related risk is studied because those risks were never attributed as the reason for any of the shutdowns, according to CEA s daily outage reports. Most of the water withdrawal and consumption factors we used are Indian median values from Chaturvedi et al. (217) because an individual power plant does not disclose its actual water-use data publicly, and no consistent database was available that reports plant-specific factors with good coverage. Hydroelectric power plants are not studied in this paper because of the difficulty of quantifying water consumption with a reasonable level of accuracy using the data available to us. Captive plants are excluded from this study because of the data limitations at the plant level. 1

11 Parched Power: Water Demands, Risks, and Opportunities for India s Power Sector WATER USE, RISKS, AND OPPORTUNITIES FOR INDIA S THERMAL POWER SECTOR Water Demands Almost 9 percent of India s thermal power generation is dependent on freshwater for cooling. As its economy grows, India s demand for electricity has been increasing rapidly over the past decades (CEA 217). Figure 1 shows how much electricity India s fossil, nuclear, and hydropower sectors have been generating every day between 213 and 216. Nuclear and hydro generation has been fairly stable on an annual basis, despite the strong seasonality of available water to produce electricity that one can observe in the hydropower sector. In contrast, fossil fuel generation has been on the rise, as shown in Figure 1. Taking a closer look at the thermal (fossil and nuclear) power sector, as illustrated in Figure 2, we found that forced outages are consistently happening throughout the year, and plants are often (61 percent of the time) generating less than what they had planned or been programmed to do in their annual generation targets. According to CEA s daily outage reports, there are many reasons for forced outages, including fuel shortages, mechanical problems, lack of available cooling water, and others. More discussion of this can be found in Section 3.2. In 216, thermal (fossil and nuclear) electricity accounted for more than 83 percent of India s total utility power generation (CEA 217), and almost 9 percent of that depended on freshwater for cooling, as shown in Figure 3, according to our analysis. The very high dependency on freshwater resources makes the country s thermal power Figure 1 India s Daily Fossil, Hydro, Nuclear, and Total Generation between 213 and 216, Solar and Wind Excluded Fossil generation grows steadily, while hydro and nuclear generation stay relatively stable 4, 3,5 Daily Generation (GWh) 3, 2,5 2, 1,5 Fossil Fuel Total 1, Hydro Nuclear Note: Electricity generated by solar and wind accounted for about 3.3% of India s total power generation in 214 according to IRENA, but is not shown on this graph, due to data availability from CEA. Vertical white strips indicate no data on that day from CEA. Source: Data from CEA, compiled and analyzed by WRI authors. WORKING PAPER January

12 Figure 2 India s Daily Programmed and Actual Generation from Thermal Power Utilities between 213 and % of the time, programmed thermal generation couldn't be delivered due to forced power plant outages, including equipment failure, fuel shortages, water shortages, and so on Daily Generation (GWh) 3, 2,5 2, 1,5 Programmed Thermal Generation Actual Thermal Generation Thermal Generation Surplus Thermal Generation Deficit Differences Between Programmed and Actual Generation (GWh) Source: Data from CEA, compiled and analyzed by WRI authors. sector extremely vulnerable to water risks like drought and freshwater scarcity from competition with other sectors. Figure 4 is a map showing the spatial distribution of all Indian thermal power plants by its source water type and cooling technology. Table 2 lists our estimates of India s thermal power utility cooling technology distribution by source water, fuel, and cooling type in 216. Distributions are shown in percentages in both generation and capacity. For freshwater-cooled utilities, our analysis with satellite images found that 13 utilities use once-through cooling systems, which is by far the most water withdrawal intensive technology. These 13 plants account for roughly 6 percent of India s total thermal capacity in 216 and are mostly located in low water-stress areas in the northeastern part of the country. About 3 percent of India s thermal electric power capacity uses dry cooling technology, which is the least water withdrawal intensive approach compared to oncethrough cooling. These are mostly concentrated in the west and the south, where water is a scarce resource. The rest, almost 82 percent of the country s total thermal capacity, is cooled with freshwater-recirculating cooling towers that typically withdraw a lot less but consume more water than once-through systems do. Twelve power plants use seawater instead of freshwater for cooling. They make up about 8.8 percent of India s total thermal capacity and are mostly located in the states of Tamil Nadu, Gujarat, and Maharashtra. Figure India s Thermal Utility Power Generation Distribution by Water Source and Cooling Technology In 216, almost 9% of India's thermal utility power generation used freshwater for cooling 8% 6% 5% 2% 7% Source: Data from CEA, compiled and analyzed by WRI authors. Seawater Once-Through Seawater Recirculating Freshwater Dry Cooling Freshwater Once-Through Freshwater Recirculating 12

13 Parched Power: Water Demands, Risks, and Opportunities for India s Power Sector Table 2 India s Thermal Power Utility Cooling Technology Distribution in 216 SOURCE WATER, FUEL, AND COOLING TYPE % TOTAL THERMAL GENERATION IN 216 % TOTAL THERMAL CAPACITY IN 216 Freshwater coal recirculating 73.4% 69.8% Freshwater coal once-through 6.4% 5.8% Seawater coal once-through 4.6% 3.7% Seawater coal recirculating 4.6% 3.3% Freshwater gas recirculating 3.1% 6.2% Seawater nuclear once-through 1.5% 1.3% Freshwater nuclear recirculating 1.5% 1.% Freshwater other recirculating 1.3% 2.7% Freshwater gas dry cooling.9% 1.5% Freshwater nuclear once-through.8%.4% Freshwater coal dry cooling.6% 1.% Freshwater oil recirculating.4% 1.3% Freshwater other dry cooling.3%.6% Freshwater biomass recirculating.3%.4% Freshwater biomass dry cooling.1%.1% Freshwater solar recirculating.1%.1% Seawater other recirculating.%.2% Seawater oil recirculating.%.1% Freshwater oil dry cooling.%.2% Seawater gas recirculating.%.2% Note: This table includes all thermal utilities included in our inventory database: 93.8%, capacity-wise, of the source water and cooling type data is developed by analyzing satellite images, the remaining 6.4% of the data is generated with our imputation model. Source: WRI authors. WORKING PAPER January

14 Figure 4 A Map of India s Thermal Power Plants by Cooling Water Source and Cooling Technology Type Delhi Ahmedabad Kolkata Mumbai Hyderabad Sources: WRI authors. Disclaimer: This map is for illustrative purposes and does not imply the expression of any opinion on the part of WRI concerning the legal status of any country or territory or concerning the delimitation of frontiers or boundaries. Freshwater Dry Cooling Freshwater Once-Through Freshwater Recirculating Seawater Once-Through Seawater Recirculating Captive Plants 14

15 Parched Power: Water Demands, Risks, and Opportunities for India s Power Sector While generation from India s thermal utilities grew by 4 percent between 211 and 216, their freshwater consumption increased by 43 percent, from 1.5 billion cubic meters in 211 to 2.1 billion in 216. According to our analysis, freshwater consumption by thermal power plants in India has been rising over the past six years, from 1.5 billion cubic meters in 211 to 2.1 billion in 216, an increase of 43 percent, as shown in Figure 5. We believe there are two primary factors driving the increase: (1) the steady growth in electricity generation, an increase of 4 percent between 211 and 216, and (2) the increased share of recirculating cooling systems, which are more water consumptive than other technologies, as illustrated in Figure 6. It is important to distinguish here between consumption water that is kept within the plant s cooling towers or evaporated to the atmosphere and withdrawal, of which a large portion (as much as 99 percent, depending on cooling technology and fuel type) may be returned to rivers and lakes and become available for use again downstream. Figure 5 India s Annual Thermal Utility Generation, Freshwater Consumption, and Withdrawal between 211 and 216 Thermal utility generation grew by 4% between 211 and 216 Freshwater consumption increased by 43% between 211 and 216 Freshwater withdrawal stayed relatively stable between 211 and 216 Terawatt-Hours 1,2 1, Billion Cubic Meters Billion Cubic Meters Sources: WRI authors; CEA (217). Figure 6 Share of Cooling Technology of Freshwater-Cooled Thermal Utilities by Installed Capacity from 211 through 216 Share of recirculating cooling among freshwater-cooled thermal utilities has increased 5% between 211 and 216 1% 3.6% 3.8% 3.9% 3.7% 3.7% 3.7% Share of Cooling Technology by Capacity 95% 9% 85% 8% 75% 1.9% 85.4% 9.7% 8.6% 86.5% 87.4% 7.8% 88.4% 7.2% 7% 89.% 89.3% 7% Source: WRI authors Recirculating Once-Through Dry WORKING PAPER January

16 While consumption has been on the rise, total freshwater withdrawal by India s thermal utilities has remained fairly stable with very small fluctuation (less than 5 percent) in the past six years, and was roughly 18.8 billion cubic meters in 216. The reason is that once-through plants withdraw water at a rate of 5 times or higher compared to other types of plants, accounting for almost 85 percent of total power sector withdrawals in India, and no new freshwater once-through plant has been introduced after 211, according to Platts. The small fluctuations in total water withdrawal from one year to another can be mainly attributed to the variation in actual generation of all once-through power plants. Even though the number of recirculating plants is increasing, the amount of additional water withdrawal from the new recirculating plants is much smaller compared to the changes in withdrawal due to the fluctuation in oncethrough plant generation. However, our estimates could be on the lower end because our method does not account for water withdrawn and consumed when power plants are not generating electricity. In the United States, more than 3 percent of all thermal power plant water withdrawal occurs when the plants are not generating electricity (Clement et al. 217), particularly for peak load plants. Cooling systems might be kept running to maintain dispatchability. Additionally, India s thermal power sector had an estimated freshwater consumption intensity of 2.2 m³/ MWh, and a withdrawal intensity of 18.9 m³/mwh, at the portfolio level (i.e., across all utilities) in 216. The key factors determining these numbers include cooling water source, cooling technology share, and power mix. Cooling water source is relatively straightforward. Plants that use seawater to cool do not consume any freshwater, but others do. Cooling systems can be generally grouped Figure 7 Statewide Average Freshwater Withdrawal Intensity vs. Consumption Intensity of the 15 Largest Thermal Electricity Producing States Some states are much more freshwater efficient than others when generating thermal electricity State-level Freshwater Consumption Intensity (m3/mwh) Telangana Chhattisgarh Andhra Pradesh Gujarat Tamil Nadu Bihar Haryana Maharashtra States with higher shares of generation cooled by seawater Odisha Madhya Pradesh Jharkhand Rajasthan Uttar Pradesh Karnataka West Bengal States with higher shares of generation cooled by once-through systems State-Level Freshwater Withdrawal Intensity (m3/mwh) Note: Bubble size denotes thermal generation for each state relative to others in 216. Source: Data from CEA and Platts, analyzed by WRI authors. 16

17 Parched Power: Water Demands, Risks, and Opportunities for India s Power Sector into three categories, once-through, recirculating, and dry cooling. Each has different water withdrawal and consumption rates because of their different heat transfer processes. Different kind of fuels have different thermal efficiencies. The more thermal-efficient a fuel is, the less waste heat per unit of generation it produces, thus the less water it needs for cooling. Different regions or states in India have different priorities when it comes to determining those three factors for power projects. For example, Gujarat is very dry but has long coastlines, so seawater cooling is used more extensively. In contrast, West Bengal is much more water abundant, thus freshwater once-through plants are feasible there. Figure 7 visualizes the statewide average freshwater withdrawal intensity against consumption intensity for the 15 largest power producing states in India. As illustrated in Figure 7, states like Gujarat and Tamil Nadu both have a high share of electricity generated by seawater-cooled plants, resulting in low intensity levels in both freshwater withdrawal and consumption. On the other end of the spectrum, states like West Bengal, Karnataka, Uttar Pradesh, and Rajasthan have relatively high shares of their thermal electricity generated by oncethrough plants, making them extremely high in terms of withdrawal intensity at the portfolio level. Risks Among all India s freshwater-cooled thermal utilities, 39 percent of its capacity is installed in high waterstress regions, generating 34 percent of its generation. Understanding the thermal power sector s water constraints at the state and national level is important because most economic and regulatory decisions are made at this level. At the same time, it is important to understand water demand and supply at the watershed level, including water demands from other sectors, because these dynamics have a substantial bearing on potential electricity generation capacity. Water flows across administrative boundaries, and upstream water-use activities, have implications for downstream user access to water despite state or national boundaries. Existing power plants could suffer from decreased water supply from increased upstream irrigated agricultural activities and might affect water supply for nearby downstream cities or villages and limit their population and economic growth. Therefore, we further examined India s thermal power plants at the watershed level using the WRI Aqueduct Global Water Risk Atlas. Figure 8 is a map of all freshwater-cooled thermal utilities against Aqueduct s Baseline Water Stress metric (Gassert et al. 214), a risk metric that measures the ratio of water demand over supply to reflect the level of competition in a watershed. The figure also includes a bar chart illustrating installed capacity distribution by water-stress level by state. For any catchment, if the water demand and supply ratio is over 4 percent meaning more than 4 percent of available water is needed and withdrawn for human use it would be considered in high water stress, which is typically the threshold we recommend to use when identifying waterstress hotspots. Among all freshwater-cooled plants, in 216, about 38.9 percent of the total generating capacity across India was installed in high (or extremely high) water-stress regions. However, these plants only generated 33.6 percent of the total freshwater-cooled thermal generation. WORKING PAPER January

18 Figure 8 India s Freshwater-Cooled Thermal Utilities Mapped against Baseline Water Stress and Distribution in Installed Capacity by Water Stress Level by State Delhi Freshwater Dry Cooling Freshwater Once-Through Ahmedabad Kolkata Freshwater Recirculating Low (<1%) Low to Medium (1 2%) Mumbai Medium to High (2 4%) Hyderabad High (4 8%) Extremely High (>8%) Arid & Low Water-Use Bangalore Chennai Maharashtra Uttar Pradesh Gujarat Madhya Pradesh Chhattisgarh West Bengal Andhra Pradesh Rajasthan Odisha Tamil Nadu Jharkhand Installed Capacity (GW) Karnataka Haryana Telangana Punjab Bihar Delhi Tripura Assam Kerala Note: Symbol size reflects the power plant s relative installed capacity. Source: WRI authors. Disclaimer: This map is for illustrative purposes and does not imply the expression of any opinion on the part of WRI concerning the legal status of any country or territory or concerning the delimitation of frontiers or boundaries. 18

19 Parched Power: Water Demands, Risks, and Opportunities for India s Power Sector India lost 14 terawatt-hours of potential thermal generation due to water shortages in 216, canceling out more than 2 percent of its growth in total electricity generation from 215. Thermal utilities around India have been suffering from forced shutdowns due to water shortages. We collected and analyzed daily outage reports published by CEA, which disclosed daily which generating unit of which power plant was out as well as the cause of the shutdown. We found that from 213 through 216, India s thermal power sector lost roughly 3 TWh in potential power generation purely due to water shortages, as shown in Figure 9. Potential power generation is calculated by multiplying the installed capacity that is out by the duration (in hours) of the outage. CEA groups outages into two categories: planned and forced. As illustrated in Figure 9, planned outages include maintenance, refurbishment, and other planned activities, accounting for about 13 percent of all outages (in TWh) between 213 and 216. Forced outages include reserve shutdowns, which are defined by CEA and include outages caused by threat to grid security, low demand, transmission congestion, and other anticipated reasons. Other forced outages include outages caused by equipment failure, fuel shortages, uneconomical operations, water shortages, and other unanticipated shutdowns. Between 213 and 216, water shortage was the fifth most frequent reason for forced outages of Indian thermal power plants and caused almost 2 percent of all outages. In 216 alone, water shortage induced potential thermal generation losses canceled out more than 2 percent of India s growth in total electricity generation between 215 and 216, as shown in Figure 1. Figure 9 A Breakdown of Planned and Forced Outages of India s Thermal Power Sector by Outage Cause between 213 and 216 Water shortage is the fifth most frequent reason for forced outages of Indian thermal power plants, accounting for 2% of all outages between 213 and Planned Outages 721 Forced Outages Total Potential Electricity Generation Losses in 4 Years (TWh) Planned Maintenance 37 Planned Refurbishment 7 Other Planned Shutdowns Equipment Failures 379 Reserve Shutdown (RSD) 174 Fuel Shortages 35 3 Uneconomical Operations Water Shortages 62 Other Unplanned Shutdowns Causes of Thermal Power Plant Outages in India Source: Data from CEA, compiled and analyzed by WRI authors. WORKING PAPER January

20 Figure 1 Comparing Water Shortage Induced Potential Thermal Power Generation Losses with the Growth of Gross Electricity Generation in India in 216 Potential generation losses due to water shortages canceled out more than 2% of India's growth in total electricity generation in 216 TWh Most of the water shortage induced outages occurred between April and September of the four years (from 213 through 216), as illustrated in Figure 11. These outages were largely driven by low water availability in the summer and delayed monsoons. Additionally, the outages were primarily concentrated in water-stressed states like Maharashtra, Gujarat, and Karnataka. A breakdown table of water shortage induced potential losses in generation, by state, can be found in the appendix. CEA s daily outage reports provided real evidence and enabled us to conduct a quantitative assessment of the severity of the impact from water shortages on India s thermal power sector. However, we did find some inconsistency and ambiguity in the outage reasons reported in CEA s reports. For example, in some instances we found that a unit was recorded as having to shut down due to cooling water pump problems, according to CEA, but, according to news reports, was offline because of water unavailability. To further understand the possible impact of water shortages on thermal generation, we analyzed capacity factors across India s thermal power generation portfolio. Growth of Gross Electricity Generation in India from 215 to 216 Water Shortages Induced Potential Thermal Power Generation Losses in 216 Source: Data from CEA, compiled and analyzed by WRI authors. Figure 11 Total Water Shortage Induced Losses in Potential Electricity Generation between 213 and 216, by Month From 213 through 216, India lost about 3 TWh in potential thermal electricity generation purely due to water shortages, mostly in months between April and September Potential Electricity Generation (TWh) Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Total Water Shortage Induced Losses Between 213 and 216 by Month Source: Data from CEA, compiled and analyzed by WRI authors. 2